ChainThink report, on March 15, the "3·15" Evening Gala exposed the rampant issue of AI large models being "poisoned." Professor Li Fumin from the Institute for Intelligent Social Governance at Shandong University of Finance and Economics stated that when merchants conduct targeted training of large models via services such as GEO, guiding AI to generate specific product or service recommendations, this constitutes a novel form of unfair competition and consumer deception through covert marketing and fabrication of facts using technical means. Consumers are unknowingly exposed to embedded promotional content, making the severity and illegality of such practices warrant serious attention. On one hand, these actions infringe upon consumers' rights to information disclosure and fair transaction as stipulated under the Consumer Rights Protection Law; on the other hand, they involve false or misleading commercial promotion through technical manipulation, disrupting legitimate recommendation algorithm systems and competitive market environments, thus constituting unfair competition.

Addressing AI poisoning requires a multi-pronged approach: regulators must prioritize monitoring AI-driven marketing inducements and strengthen enforcement oversight; AI operators should enhance scrutiny of training data sources and implement output filtering mechanisms with traceability; consumers must improve their ability to identify the commercial nature of AI-generated content and actively exercise their rights through complaints and reports. (Xinhua News Agency)